A multi-view genomic data simulator.

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Omics technologies enable large-scale feature analysis (e.g., mRNA, miRNA expression, methylation).
• Key challenge in computational method development is the lack of annotated biological datasets for benchmarking.
• Generating synthetic datasets offers a controlled environment for method evaluation.

Purpose of the Study:

• To propose a novel method for generating synthetic biological datasets.
• To create realistic datasets for benchmarking feature selection and data mining techniques.
• To facilitate the development and validation of computational tools in omics research.

Main Methods:

• Generation of interaction networks among biological molecules, particularly those regulating gene expression.
• Utilizing ordinary differential equations (ODEs) with known parameters to derive synthetic datasets.
• Ensuring synthetic data mimics the behavior of real biological data.

Main Results:

• Generated synthetic datasets effectively mimic real omics data.
• The method allows for the selective identification of molecular interactions.
• Popular data analysis methods successfully identify known interactions within the synthetic data.

Conclusions:

• The proposed method aids in assessing data mining techniques alongside real biological datasets.
• Key strength lies in full control over simulated data while maintaining biological relevance.
• An R package, MVBioDataSim, is available for community use.